FarmAnimal Digital Twin Framework for Predictive Health Monitoring and Welfare Improvement

Abstract

In the context of growing climatic, economic and regulatory stresses, livestock production systems must continue to protect, promote and maintain animal health and welfare while maintaining productivity. The current health surveillance in farm environments is mostly based on a periodic visual check and a "proactive" approach to veterinary intervention when the results of the check are not satisfactory, which are both labour-consuming and subjective, and often too late to stop the development of animal disease or distress. This paper introduces the FarmAnimal Digital Twin (FADT) framework, which is a conceptual and computational model that builds virtual copies of individual farm animals from multiple streams of data collected from sensors, animals and environments that evolve in real time. The framework is based on a combination of a data acquisition layer and a synchronization and modelling layer along with a predictive analytics layer that allows for early detection of anomalies, personalized health scores, and welfare-based decision support. The proposed digital twin is a persistent virtual state of the animal that evolves in real time through the incoming physiological, locomotor and environmental signals that enables predictive rather than descriptive assessment of a virtual state of the animal. The framework is defined by the functional layers, information flow, and the assessment criteria, and is located in the context of current digital twin applications in dairy, poultry, swine, and cattle production. A tabular comparative analysis of the monitored parameters, predictive tasks and welfare indicators is provided to illustrate the range and the applicability of welfare framework across different species and production systems. The discussion covers the issues of data interoperability, model generalizability and challenges faced in the adoption of models by small- and medium-sized farmers. The FADT framework is designed to inform and inspire future implementation and empirical validation research to measure the benefits that the FADT can bring in terms of early diagnosis of disease, avoidance of antimicrobial use, and better overall animal welfare outcomes in various farming systems.